biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Core] Support reseting all running requests' KV while calling reset_prefix_cache (#28827)

Browse Source 
Signed-off-by: Zhuohan Li <zhuohan123@gmail.com>
Signed-off-by: Nick Hill <nhill@redhat.com>
Co-authored-by: Nick Hill <nhill@redhat.com>
This commit is contained in:
Zhuohan Li
2025-12-01 18:25:05 -08:00
committed by GitHub
parent fa8804ad9c
commit d0cd728907
16 changed files with 315 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
vllm/entrypoints/openai/api_server.py
View File

@@ -877,13 +877,15 @@ if envs.VLLM_SERVER_DEV_MODE:
return JSONResponse(content=server_info)
@router.post("/reset_prefix_cache")
async def reset_prefix_cache(raw_request: Request):
async def reset_prefix_cache(
raw_request: Request, reset_running_requests: bool = Query(default=False)
):
"""
Reset the prefix cache. Note that we currently do not check if the
prefix cache is successfully reset in the API server.
"""
logger.info("Resetting prefix cache...")
await engine_client(raw_request).reset_prefix_cache()
await engine_client(raw_request).reset_prefix_cache(reset_running_requests)
return Response(status_code=200)
@router.post("/reset_mm_cache")